Fungicide impacts on microbial communities in soils with contrasting management histories

Chemosphere 69 (2007) 82–88 www.elsevier.com/locate/chemosphere

Fungicide impacts on microbial communities in soils with contrasting management histories Gary D. Bending *, M. Sonia Rodrı´guez-Cruz 1, Suzanne D. Lincoln Warwick HRI, University of Warwick, Wellesbourne, Warwick CV35 9EF, UK Received 7 December 2006; received in revised form 7 March 2007; accepted 13 April 2007 Available online 1 June 2007

Abstract The impacts of the fungicides azoxystrobin, tebuconazole and chlorothalonil on microbial properties were investigated in soils with identical mineralogical composition, but possessing contrasting microbial populations and organic matter contents arising from different management histories. Degradation of all pesticides was fastest in the high OM/biomass soil, with tebuconazole the most persistent compound, and chlorothalonil the most readily degraded. Pesticide sorption distribution coefficient (Kd) did not differ significantly between the soils. Chlorothalonil had the highest Kd (97.3) but Kd for azoxystrobin and tebuconazole were similar (13.9 and 12.4, respectively). None of the fungicides affected microbial biomass in either soil. However, all fungicides significantly reduced dehydrogenase activity to varying extents in the low OM/biomass soil, but not in the high OM/biomass soil. The mineralization of subsequent applications of herbicides, which represents a narrow niche soil process was generally reduced in both soils by azoxystrobin and chlorothalonil. 16S rRNAPCR denaturing gradient gel electrophoresis (DGGE) indicated that none of the fungicides affected bacterial community structure. 18S rRNA PCR-DGGE analysis revealed that a small number of eukaryote bands were absent in certain fungicide treatments, with each band being specific to a single fungicide–soil combination. Sequencing indicated these represented protozoa and fungi. Impacts on the specific eukaryote DGGE bands showed no relationship to the extent to which pesticides impacted dehydrogenase or catabolism of herbicides. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Fungicides; Denaturing gradient gel electrophoresis; Dehydrogenase; Microbial community structure; Eukaryotes; Soil management

1. Introduction Since pesticides can potentially exert toxicity to organisms other than their intended target, there has been considerable interest in determining their impacts on nontarget organisms within the soil, and the implications of such effects for soil processes. Most of the literature dealing with non-target effects of pesticides on soil microbes has investigated ecotoxicological impacts on broad-scale soil * Corresponding author. Tel.: +44 (0) 24 76575057; fax: +44 (0) 24 7657 4500. E-mail address: [email protected] (G.D. Bending). 1 Present address: Institute of Natural Resources and Agrobiology (IRNASA-CSIC), Department of Environmental Chemistry and Geochemistry, Apdo. 257, Salamanca 37071, Spain.

0045-6535/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2007.04.042

properties, to which all soil microbes contribute (Bu¨nemann et al., 2006), including rates of respiration (e.g. Chen et al., 2001) and amounts of microbial biomass (e.g. Hart and Brookes, 1996; Smith et al., 2000). Such broad-scale approaches form the basis of OECD tests used by regulatory authorities to measure impacts of pesticides on soil microbes (OECD, 2000). While broad-scale approaches to soil microbial analysis provide an indication of net community responses, they could potentially mask community changes which may be more subtle, but nonetheless of ecological significance. In particular, broad-scale approaches to measuring pesticide impacts take no account of microbial diversity, which may be a crucial contributor to soil quality, controlling long-term sustainability and resistance to perturbations (Lynch et al., 2004). Furthermore, broad-scale processes

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

may respond differently to perturbations than narrow niche soil processes driven by specific microbial groups, for which there may be less functional redundancy (Girvan et al., 2005). The extent to which pesticide impacts on broad-scale microbial community properties and processes can be used to predict potential impacts on such fine scale properties and processes is unclear. Relatively few studies have used PCR-based community profiling methods such as denaturing gradient gel electrophoresis (DGGE) and terminal restriction fragment length polymorphism (TRFLP) to investigate ‘fine-scale’ impacts of pesticides on soil microbial communities. Several studies using PCR-DGGE or TRFLP have shown that herbicides can have impacts on the structure of soil bacterial communities (Engelen et al., 1998; Chang et al., 2001; Rousseaux et al., 2003). In contrast, while fungicides have commonly been shown to impact broad-scale microbial properties and activities (Pal et al., 2005; Bu¨nemann et al., 2006), few studies have used culture-independent methods to investigate impacts on the structure and diversity of microbial communities. Using DGGE, Sigler and Turco (2002) found that the fungicide chlorothalonil impacted bacterial and fungal communities, although the longevity of the effects, the extent to which impacts were associated with compound persistence, and impacts on broad-scale microbial properties, were not determined. Impacts of perturbations on specific microbial communities may reflect direct impacts, or indirect effects via organisms inhabiting higher or lower trophic levels. However, those studies in which pesticide impacts on microbial community structure have been investigated using cultureindependent methods have generally focussed on bacterial or fungal communities alone, and have not considered other important soil microbial groups such as protozoa and oomycetes. Furthermore, the degree to which pesticides impact soil microbial communities, including the magnitude of the impact and the rate and extent of recovery, will reflect a variety of interacting biotic and abiotic factors. A range of inherent soil characteristics such as size and structure of the microbial community, and organic matter content, are likely to control pesticide impacts on soil microbial communities by determining exposure through biodegradation and bioavailability. However these relationships remain to be elucidated. The aim of the current study was to use soils from the same soil series, but with contrasting microbial populations and soil organic matter (SOM) content resulting from different management histories, to investigate the impacts of fungicides on soil microbial communities. Using three fungicides with contrasting modes of action we analysed impacts on broad- and fine-scale microbial properties to address the following questions: (1) What is the relative susceptibility of broad and fine scale microbial characteristics to fungicide impacts? (2) Are the impacts of fungicides on broad and fine scale microbial community properties linked? (3) How does management pre-history affect the responses of microbial communities to fungicides?

83

2. Materials and methods 2.1. Soil collection and analysis Soil samples were collected from sites under conventional and organic management on the farm at Warwick HRI, Wellesbourne, Warwickshire, UK, during November 2003. In both fields the soil is a sandy loam of the Wick series, with 73% sand, 12% silt and 14% clay (Whitfield, 1974). Soil was taken from Long Close field, which had been under conventionally managed cereal rotation for at least 20 years. Fungicides applied over this time included applications of tebuconazole in 1999 and 2001, and chlorothalonil in 2001. Soil was also collected from Hunts Mill field, which had been converted from a conventional cereal rotation to an organically managed grass-clover ley 10 years prior to soil collection. The organic field had received no pesticide or fertiliser since conversion. Within each field, 9 kg of top-soil (0–10 cm) was collected from four locations along a 20 m transect. Methods involved in soil collection, processing and characterisation of C, N and pH were as described in Bending et al. (2006). 2.2. Soil–fungicide incubation Suspensions of the commercial formulations of the fungicides tebuconazole, (((RS)-1-p-chlorophenyl-4,4-dimethyl-3-(1H-1,2,4-triazol-1-ylmethyl)pentan-3-ol), Bayer CropScience, Cambridge, UK), azoxystrobin, ((methyl (E)-2-{2-[6-(2-cyanophenoxy)pyrimidin-4-yloxy]phenyl}-3methoxyacrylate), Syngenta Crop Protection, Bracknell, UK) or chlorothalonil (tetrachloroisophthalonitrile, Atlas Crop Protection Ltd., Doncaster, UK) in H2O were added to separate 600 g fresh weight portions of soil from each location to give concentrations of 5 mg kg 1 (tebuconazole and azoxystrobin) or 10 mg kg 1 (chlorothalonil). The concentrations used reflect the maximum recommended fungicide doses dispersed in the top 1 cm of soil. The fungicides were chosen to represent a range of solubilities and capacities to sorb to soil organic matter, with organic-carbon normalised sorption coefficient (Koc) of 500, 906–1251 and 1600–14 000 for azoxystrobin, tebuconazole and chlorothalonil, respectively (Tomlin, 2003) and hence contrasting potential bioavailabilities. The fungicides have contrasting modes of action, with tebuconazole inhibiting ergosterol biosynthesis, chlorothalonil inhibiting conjugation of thiols, and azoxystobin blocking mitochondrial respiration (Tomlin, 2003). Additional H2O was added to bring the soil water holding capacity (WHC) to 40%, which was approximately 33 kPa. The fungicides were uniformly mixed into separate soil samples as described by Bending et al. (2006). Control soils were also set up which received H2O without fungicide. Amended soils were incubated in polypropylene bottles at 15 °C. Soil moisture content was maintained by addition of sterile distilled water as necessary to keep the soil at a constant WHC of 40%.

84

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

2.3. Fungicide analysis Total fungicide remaining in the soils, and the portion in the soil solution, was determined at time 0, and after 7 days and 1, 2 and 3 months. Methods for extraction of total and soil solution fungicide, and analysis of fungicide concentrations by HPLC were as described in Bending et al. (2006), with detection by UV absorbance at 220 nm for tebuconazole and azoxystrobin and 230 nm for chlorothalonil. The sorption distribution coefficient (Kd) was determined by calculating the proportion of fungicide in the soil solution relative to that remaining in the soil (Walker et al., 1999). 2.4. Effects of fungicides on broad-scale soil microbial community properties After 7 days, 1, 2 and 3 months, biomass-N and dehydrogenase, to which all living organisms within the soil contribute, and which represent broad-scale measurements of the amount and activity of soil organisms, respectively, were measured in control (untreated) and fungicide treated soils. Microbial biomass-N was extracted using the chloroform fumigation-extraction technique (Joergensen and Brookes, 1990). Ninhydrin-reactive N released by fumigation was converted to biomass-N using a conversion factor of 3.1 (Amato and Ladd, 1988). Dehydrogenase activity was measured using the method of Tabatabai (1994). Basal respiration was determined at time 0 by measuring CO2 evolution from 20 g portions of soil incubated at 25 °C for 1 h, using an infra red gas (ADC, Hoddesdon, Surrey, UK). 2.5. Effects of fungicides on degradation of subsequent herbicides Degradation of pesticides is a key function of agricultural soil which reduces contamination of the wider environment. Pesticide degradation is typically mediated by specific soil microbes possessing suitable catabolic enzymes (Aislabie and Lloyd-Jones, 1995), and therefore represents a narrow niche function. The effect of previous fungicide application on degradation of the herbicides isoproturon (3-(4-isopropylphenyl)-1,1-dimethylurea, Atlas Agrochemicals, Kent, UK) or bentazone (3-isopropyl-1H-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide, BASF, Suffolk, UK) was determined after 3 months, to determine whether there had been any medium-term impacts of the fungicides on the attenuation capacity of the soils. [ring-U-14C]Isoproturon (>99% purity), and [carbonyl-14C]bentazone (>95% purity) were supplied by Bayer Corp. (Germany) and Izotop Co. (Hungary), respectively. 10 g fw of control, tebuconazole, azoxystrobin or chlorothalonil amended soil were placed in 250 ml Duran bottles and 30 ml herbicide formulation solution spiked with 14C-ring labelled analogue was added. The final concentration was 5 lg ml 1 herbicide and the activity was 100 Bq ml 1. 14CO2 evolved was collected and measured after 2, 4, 7 and 10 days using

the method described by Reid et al. (2001), and cumulative herbicide mineralization calculated. 2.6. Effects of fungicides on the structure of soil prokaryote and eukaryote communities After 3 months, DNA was extracted from 1 g fw portions of soil by bead-beating using a Cambio Ultraclean soil DNA extraction kit. The DNA was amplified using bacteria or eukaryote specific small subunit (SSU) rRNA gene primers and the community profiled using denaturing gradient gel electrophoresis (DGGE). Eukaryote communities were amplified using the 18S rRNA primers F1427GC and R1616 (van Hannen et al., 1998), and bacteria were amplified using primers corresponding to positions 341–534 of E. coli, as described by Muyzer et al. (1993). PCR reaction mixture was as described by Bending et al. (2003). In order to determine the nature and diversity of soil eukaryotes amplified by primers F1427GC and R1616, a sample of DNA extracted from unamended soil from Long Close field was amplified with the primers, as described above. The PCR products were purified using a QIAquick PCR purification kit (Qiagen Ltd., Dorking, UK), and then cloned using a TOPO cloning kit (Invitrogen, Paisley, UK). Plasmids were extracted from 25 clones containing an insert using a Qiagen Plasmid extraction kit. Plasmid DNA was used in sequencing reactions conducted according to Bending et al. (2003). DGGE gels were set up and visualised according to Bending and Rodriguez-Cruz (2007). A number of eukaryote bands showed consistent reduction in intensity or loss in all four replicates of certain treated soils relative to the corresponding untreated soil, and these were cut from the gel, and DNA amplified as described in Bending et al. (2003). Re-amplified bands were run against the original sample to check motility and purity, before cloning and sequencing as described above. For each band 12 clones were sequenced, and sequences appearing more than twice in the clone library selected for further analysis. The partial 18S rRNA sequences were edited and assembled using the DNAstar II sequence analysis package (Lasergene Inc., Madison, Wisconsin, USA). Sequences were compared to those on the EMBL DNA database using the program BLAST. The GelCompar II package (Applied Maths, Ghent, Belgium) was used to analyse DGGE fingerprints, in order to determine pesticide impacts on community structure. Bands were counted if they comprised greater than 1% of the total peak area. Presence and absence of bands was used for comparison of fungicide effects on bacterial and fungal community structure using principle component analysis (PCA). 2.7. Statistical analysis Significance of differences between treatments were determined by analysis of variance (ANOVA) using the program GenStat (7th Edition, VSN International Ltd.).

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

70.8% and 40.2% remaining in Long Close and Hunts Mill soils, respectively, after 3 months. Chlorothalonil was metabolised relatively quickly, with 8.6% and 0.4% remaining in Long Close and Hunts Mill soil after 3 months. There was no significant difference in sorption of any of the fungicides between Long Close and Hunts Mill soil. Average Kd values one day following addition to soil were 12.4, 13.9 and 97.3 for tebuconazole, azoxystrobin and chlorothalonil, respectively. Kd of all the fungicides increased over time as the proportion of the pesticide in the soil solution declined, with solution phase fungicide concentrations falling to below limits of detection within 1 month for tebuconazole and chlorothalonil, and 2 months for azoxystrobin (data not shown).

3. Results 3.1. Soil characteristics While Long Close and Hunts Mill soils had comparable pH, Long Close had approximately 25% lower organic C and N contents than Hunts Mill (Table 1). Biomass-C, dehydrogenase and basal respiration were, respectively, over 4, 6 and 3 times higher in Hunts Mill relative to Long Close soil (Table 1). 3.2. Persistence of fungicides The degradation rate of all 3 fungicides was faster in Hunts Mill relative to Long Close soil. Tebuconazole degraded slowly, with 83.7% and 71.5% remaining in Long Close and Hunts Mill soils, respectively, after 3 months (Fig. 1). Azoxystrobin was moderately persistent, with

3.3. Effects of fungicides on broad-scale soil microbial community properties Fungicide addition had no significant effect on microbial biomass in either soil at any time point (data not shown). None of the fungicides affected dehydrogenase in Hunts Mill soil (data not shown). In Long Close soil, dehydrogenase activity was significantly reduced (P < 0.05) in the azoxystrobin treatment by between 10.7% and 13.7% during the first month following application, but after 2 months, activity was no different to that in the control (Table 2). Chlorothalonil significantly reduced (P < 0.05) dehydrogenase by between 23.0% and 39.1% in Long Close soil, with no recovery within a 3 month timescale. At the 1 month time interval only, tebuconazole significantly reduced dehydrogenase activity by 10.0% (P < 0.05).

Table 1 Chemical and microbial characteristics in Long Close and Hunts Mill soils prior to fungicide addition

pH Total organic C (%) Total organic N (%) Biomass (mg C kg 1 dw soil) Dehydrogenase (lg TPF g 1 dw soil) Basal respiration (ll CO2 g dw soil h 1)

Long Close

Hunts Mill

6.5 (0.04) 1.18 (0.03) 0.09 (0.01) 139.4 (53.9) 56.0 (3.3) 10.8 (1.2)

6.5 (0.07) 1.62 (0.01) 0.12 (0.01) 622.6 (83.9) 349.7 (48.4) 39.8 (3.6)

Data represent mean of four replicates. Figures in brackets are +/ standard error of the mean.

Azoxystrobin % Azoxystrobin remaining

% Tebuconazole remaining

Tebuconazole 100 80 60 40 20 0 0

20

40

60

85

80

100

100 80 60 40 20 0 0

20

Time (days)

40

60

80

100

Time (days)

% Chlorothalonil remaining

Chlorothalonil 100 80 60 40 20 0 0

20

40

60

80

100

Time (days) Fig. 1. Recovery of fungicide: (.) Hunts Mill soil; (d) Long Close soil. Bars represent +/

standard error of the mean.

86

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

Table 2 Effect of fungicides on soil dehydrogenase activity in Long Close soil. Data represents mean of four replicates Dehydrogenase activity (as % of unamended control) Time (months) Azoxystrobin Tebuconazole Chlorothalonil a

0.25 89.4a 97.0 77.0a

1 86.3a 90.0a 62.6a

2 97.7 96.7 68.2a

3 113.7 111.6 60.9a

Indicates significant difference to respective control (P < 0.05).

3.4. Effects of fungicides on mineralization of subsequent herbicide additions Effects of pre-treatment with fungicides on amounts of a subsequent herbicide addition which had been mineralized after 10 days were similar in the 2 soils (Table 3). Bentazone mineralization was significantly inhibited (P < 0.05) only by chlorothalonil, which reduced degradation by 36.0% and 16.4% in Long Close and Hunts Mill soils, respectively. Chlorothalonil also significantly reduced (P < 0.05) isoproturon mineralization, which declined by 19.7% in Long Close soil, and 13.7% in Hunts Mill soil. Azoxystobin significantly reduced (P < 0.05) mineralization of isoproturon by 10.8% and 6.9% in Long Close and Hunts Mills soils, respectively. 3.5. Effects of fungicides on prokaryote and eukaryote community structure The eukaryote primers were shown to amplify DNA from a wide range of eukaryotes (data not shown). Fungi were represented by 16 of the 25 sequenced clones, with eight ascomycete fungi, five Zygomycetes related to Mortierella sp., and three basidiomycetes. Protozoa formed the second largest group, with four flagellates, and single Ciliate, Icthyosporea and Amoebae clones. There were single nematode and Chytridiomycete (oomycete) clones. An average of 30 and 27.5 18S rRNA-DGGE bands were recorded in Hunts Mill and Long Close soils, respectively, although band number was not significantly affected by fungicide application. However, fungicide treatment affected a small number of specific 18S rRNA-DGGE bands in both soils (Fig. 2a and b). Three bands which were present in control soil but were absent or their intensity severely reduced in all fungicide treated replicates were Table 3 Effect of pre-treatment of Long Close and Hunts Mill soils with fungicides for 3 months on mineralization of a subsequent herbicide addition during a 10 day incubation. Data represents mean of 4 replicates Fungicide

Mineralization of herbicide (as % of unamended control) Long Close

Tebuconazole Azoxystrobin Chlorothalonil a

Fig. 2. 18S rRNA PCR-DGGE profile in control and fungicide treated soils after 3 months. 1–4 control; 5–8 tebuconazole; 9–12 azoxystrobin; 13–16 chlorothalonil. Positions of sequenced bands indicated. (a) Long Close, (b) Hunts Mill.

cut from 18S rRNA-DGGE gels, cloned and sequenced. In Long Close soil, band Lch which was absent from chlorothalonil treated soil was shown to represent ciliate protozoans, with equal numbers of clones showing homology to Arcuospathidium sp. DQ411860, and to Breslaua vorax AF060453 (Table 4). In Hunts Mill soil band HM1, which was absent from axoxystrobin treated soil showed closest sequence homology to the flagellate protozoan Paraflabellula hoguae AY277797. Band HM2, which showed severely reduced intensity in tebuconazole treated soil showed greatest homology to the ascomycete fungus Cladosporium tenuissimum AJ301715. However, the degree of similarity was only 78%. Principle component analysis of 18S rRNA-DGGE profiles revealed that none of the pesticides significantly affected overall community structure in either soil (data Table 4 Closest matches of DNA cloned from excised 18SrRNA-DGGE bands to sequences from the EMBL database, and nucleotide accession numbers of cloned sequences Band

EMBL accession no.

Closest sequence match

Similarity (%)

Taxonomic affiliation

Lch

AM159607

Arcuospathidium cultriforme DQ411860 Bresslaua vorax AF060453

97

Cilliate protozoan Ciliate protozoan

Paraflabellula hoguae AY277797 Cladosporium tenuissimum AJ301715

87

Hunts Mill

AM159608

Bentazone

Isoproturon

Bentazone

Isoproturon

103.0 83.1 64.0a

100.9 89.2a 80.3a

98.0 96.7 83.6a

97.3 93.1a 86.3a

Indicates significant difference to control (P < 0.05).

HM1

AM159611

HM2

AM159617

99

78

Flagellate protozoan Ascomycete fungus

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

not shown). There were no fungicide induced changes to bacterial DGGE banding in either soil. 4. Discussion While responses of biomass and catabolic potential to the fungicides were the same in the soils, other parameters, including dehydrogenase and specific organisms identified by 18S rRNA-DGGE responded differently in the soils. Differences in responses of the microbial properties to fungicides in the two soils could reflect a variety of factors, including differences in structure of the microbial community and the extent to which soil communities were exposed to the fungicides. The degradation rates of the fungicides were higher in Hunts Mill relative to Long Close soil, so that exposure of soil microbes to the fungicides was more prolonged in Long Close soil relative to Hunts Mill soil. All of the fungicides became tightly sorbed on addition to both soils, with Kd increasing over time, as the proportion found in the soil solution declined. This could have reflected preferential degradation of pesticide in the soil solution or increased sorption via ‘ageing’ over time. Although there were no differences in sorption of the fungicides between the two soils, sorption is generally considered a poor indicator of bioavailability. Differences in degradation rate and bioavailability in the soils could therefore have resulted in contrasting exposure, explaining why dehydrogenase was only affected by fungicides in Long Close soil, and the greater impact of fungicides on herbicide catabolism in Long Close relative to Hunts Mill soil. DGGE analysis revealed a small number of eukaryote bands which were absent or reduced in fungicide treated soils, with each band being specific to a single fungicide– soil combination. Such changes could reflect direct inhibition of organisms by the fungicides or indirect changes reflecting altered competitive interactions between microbes in the presence of the fungicide. Different impacts of fungicides on 18S rRNA banding in the soils could reflect differences in community structure, which was reflected by contrasting 18S rRNA-DGGE profiles in the two soils. However, the general 18S rRNA-DGGE primers used in the study profile dominant eukaryotes only, and group specific primers would be required to determine the full extent of impacts on specific eukaryote groups. The small number of changes in DGGE banding indicates that impacts on community structure were limited. Sigler and Turco (2002) showed that chlorothalonil removed a number of bands from the fungus community DGGE profile of agricultural and turfgrass soils 2 weeks following application. Similarly, heavy metals and benzene can reduce the number or diversity of bands in bacterial DGGE gels, indicating toxicity responses (Girvan et al., 2005; Lorenz et al., 2006). Sequencing revealed that DGGE bands impacted by chlorothalonil and tebuconazole represented protozoa in Long Close and Hunts Mill soil, respectively, while azoxystrobin affected an ascomycete fungus in Hunts Mills soil.

87

Since the fungicides had contrasting modes of action, differences in specific organisms affected by the fungicides would be expected. Similarly a number of studies using culture-dependent methods have indicated that a variety of fungicides can reduce populations of protozoa at recommended application rates, with direct rather than indirect effects responsible (Ekelund, 1999; Ekelund et al., 2000). Catabolism of xenobiotic compounds in soil is considered a narrow niche function performed by a small subset of the soil microbiota with specific metabolic capabilities (Girvan et al., 2005). Chlorothalonil and azoxystrobin reduced catabolism of subsequent herbicide additions in both soils, indicating that they had reduced the activities of the specific community involved in catabolism of these herbicides. Furthermore, these effects on herbicide catabolism were noted after 3 months exposure, at a time when azoxystrobin impacts on dehydrogenase had recovered, suggesting that the herbicide catabolising community showed lower resilience to azoxystrobin than dehydrogenase activity. Similarly, other organic and metal pollutants have been shown to have greater effects on narrow niche functions associated with catabolism of xenobiotics, relative to broad-scale soil processes (Girvan et al., 2005; Lors et al., 2005). Other narrow niche microbial functions, including nitrification and methane oxidation can also be inhibited by pesticide application (Chang et al., 2001; Seghers et al., 2003). Given the difficulties of interpreting changes in broad-scale microbial properties, or changes in the gross structure of microbial communities in terms of functioning, narrow niche functions clearly have great potential as bioindicators of ecotoxicological impacts, and for assessment of soil quality. 5. Conclusions We conclude that there was no clear relationship between the responses of the different broad and fine scale microbial properties to the pesticides and that fungicide impacts on the different microbial properties were not predictable in terms of the size of the impact or the duration of the effect. In some treatments the fungicides affected the narrow niche function of herbicide catabolism without necessarily impacting broad-scale activity measured using dehydrogenase, and vice versa. Similarly, DGGE analysis revealed impacts on specific eukaryote community members without impacts on any other parameter. Chlorothalonil generally had greater and more prolonged impacts on soil microbial properties than azoxystrobin or tebuconazole. Soil pre-history affected the responses of microbial communities to fungicides, probably via effects of differences in organic matter content and the size of the microbial community, which affected biodegradation rates and possibly also bioavailability, and which could have affected the level of exposure of soil communities. Interpretation of ecotoxicological impacts on soil using broad-scale properties can be problematic, since microbial death caused by impacts can be followed by growth of

88

G.D. Bending et al. / Chemosphere 69 (2007) 82–88

organisms using the killed biomass as a substrate. This could mask negative effects or result in net increases in the broad-scale properties, rather than reductions. Accurate assessment of the ecotoxicological impact of pesticides clearly requires assessment at a range of biological scales, and further work is required to develop methods for using microbial community profiling and measurement of narrow niche functions for impact assessment. Acknowledgements We thank the Department of Environment, Food and Rural Affairs for funding (project PL0550). M. Sonia Rodriguez Cruz thanks the Spanish Ministry of Education and Science for her postdoctoral fellowship. References Aislabie, J., Lloyd-Jones, G., 1995. A review of bacterial degradation of pesticides. Aust. J. Soil Res. 33, 925–942. Amato, M., Ladd, J.N., 1988. Assay for microbial biomass based on ninhydrin-reactive nitrogen in extracts of fumigated soils. Soil Biol. Biochem. 20, 107–114. Bending, G.D., Rodriguez-Cruz, M.S., 2007. Microbial aspects of the interaction between soil depth and biodegradation of the herbicide isoproturon. Chemosphere 66, 664–671. Bending, G.D., Lincoln, S.D., Sorensen, S.R., Morgan, J.A.W., Aamand, J., Walker, A., 2003. In-field spatial variability in the degradation of the phenyl-urea herbicide isoproturon is the result of interactions between degradative Sphingomonas spp. and soil pH. Appl. Environ. Microbiol. 69, 827–834. Bending, G.D., Lincoln, S.D., Edmondson, R.N., 2006. Spatial variation in the degradation rate of the pesticides isoproturon, azoxystrobin and diflufenican in soil and its relationship with chemical and microbial properties. Environ. Pollution 139, 279–287. Bu¨nemann, E.K., Schwenke, G.D., Van Zwieten, L., 2006. Impact of agricultural inputs on soil organisms – a review. Aust. J. Soil Res. 44, 379–406. Chang, Y.J., Hussain, A.K.M.A., Stephen, J.R., Mullen, M.D., White, D.C., Peacock, A., 2001. Impact of herbicides on the abundance and structure of indigenous beta-subgroup ammonia-oxidizer communities in soil microcosms. Environ. Toxicol. Chem. 20, 2462–2468. Chen, S.K., Edwards, C.A., Subler, S., 2001. Effects of the fungicides benomyl, captan and chlorothalonil on soil microbial activity and nitrogen dynamics in laboratory incubations. Soil Biol. Biochem. 33, 1971–1980. Ekelund, F., 1999. The impact of the fungicide fenpropimorph (Corbel (R)) on bacterivorous and fungivorous protozoa in soil. J. Appl. Ecol. 36, 233–243. Ekelund, F., Westergaard, K., Soe, D., 2000. The toxicity of the fungicide propiconazole to soil flagellates. Biol. Fert. Soils 31, 70–77. Engelen, B., Meinken, K., von Wintzintgerode, F., Heuer, H., Malkomes, H.P., Backhaus, H., 1998. Monitoring impact of a pesticide treatment on bacterial soil communities by metabolic and genetic fingerprinting in addition to conventional testing procedures. Appl. Environ. Microbiol. 64, 2814–2821. Girvan, M.S., Campbell, C.D., Killham, K., Prosser, J.I., Glover, L.A., 2005. Bacterial diversity promotes community stability and functional resilience after perturbation. Environ. Microbiol. 7, 301–331.

Hart, M.R., Brookes, P.C., 1996. Soil microbial biomass and mineralisation of soil organic matter after 19 years of cumulative field applications of pesticides. Soil Biol. Biochem. 28, 1641–1649. Joergensen, R.G., Brookes, P.C., 1990. Ninhydrin-reactive measurements of microbial biomass in 0.5 M K2SO4 soil extracts. Soil Biol. Biochem. 22, 1023–1027. Lorenz, N., Hintemann, T., Kramarewa, T., Katayama, A., Yasuta, T., Marschner, P., Kandeler, E., 2006. Response of microbial activity and microbial community composition in soils to long-term arsenic and cadmium exposure. Soil Biol. Biochem. 38, 1430–1437. Lors, C., Lagacherie, B., Chabanet, C., Soulas, G., 2005. DNOC, a model pollutant, adversely affects the potential of soil microbial communities to mineralise the herbicide 2,4-D: an investigation using microsampling procedures. Soil Biol. Biochem. 37, 1023–1032. Lynch, J.M., Benedetti, A., Insam, H., Nuti, M.P., Smalla, K., Torsvik, V., Nannipieri, P., 2004. Microbial diversity in soil: ecological theories, the contribution of molecular techniques and the impact of transgenic plants and transgenic microorganisms. Biol. Fert. Soils 40, 363–385. Muyzer, G., Dewaal, E.C., Uitterlinden, A.G., 1993. Profiling of complex microbial-populations by denaturing gradient gel-electrophoresis analysis of polymerase chain reaction-amplified genes-coding for 16S ribosomal-RNA. Appl. Environ. Microbiol. 59, 695–700. OECD, 2000. OECD guidelines for testing of chemicals. 217 Soil microorganisms: carbon transformation test. Organisation for Economic Co-operation and Development, Paris. Pal, R., Chakrabarti, K., Chakraborty, A., Chowdhury, A., 2005. Pencycuron application to soils: degradation and effect on microbiological parameters. Chemosphere 60, 1513–1522. Reid, B.J., MacLeod, C.J.A., Lee, P.H., Morriss, A.W.J., Stokes, J.D., Semple, K.T., 2001. A simple C-14-respirometric method for assessing microbial catabolic potential and contaminant bioavailability. FEMS Microbiol. Lett. 196, 141–146. Rousseaux, S., Hartmann, A., Rouard, N., Soulas, G.A., 2003. A simplified procedure for terminal restriction fragment length polymorphism analysis of the soil bacterial community to study the effects of pesticides on the soil microflora using 4,6-dinitroorthocresol as a test case. Biol. Fert. Soils 37, 250–254. Seghers, D., Bulcke, R., Reheul, D., Siciliano, S.D., Top, E.M., Verstraete, W., 2003. Pollution induced community tolerance (PICT) and analysis of 16S rRNA genes to evaluate the long-term effects of herbicides on methanotrophic communities in soil. Eur. J. Soil Sci. 54, 679–684. Sigler, W.V., Turco, R.F., 2002. The impact of chlorothalonil application on soil bacterial and fungal populations as assessed by denaturing gradient gel electrophoresis. Appl. Soil Ecol. 21, 107–118. Smith, M.D., Hartnett, D.C., Rice, C.W., 2000. Effects of long-term fungicide applications on microbial properties in tallgrass prairie soil. Soil Biol. Biochem. 32, 935–946. Tabatabai, M.A., 1994. Soil enzymes. In: Weaver, R.W. (Ed.), Methods of Soil Analysis: Part 2. Microbiological and Biochemical Properties. Soil Science Society of America, Madison, Wisconsin, USA, pp. 903–947. Tomlin, C.D.S., 2003. The Pesticide Manual, 13th ed. British Crop Protection Council, Surrey, UK. van Hannen, E.J., van Agterveld, M.P., Gons, H.J., Laanbroek, H.J., 1998. Revealing genetic diversity of eukaryotic microorganisms in aquatic environments by denaturing gradient gel electrophoresis. J. Phycol. 34, 206–213. Walker, A., Turner, I.J., Cullington, J.E., Welch, S.J., 1999. Aspects of the adsorption and degradation of isoproturon in a heavy clay soil. Soil Use Manage. 15, 9–13. Whitfield, W.A.D., 1974. The soils of the National Vegetable Research Station, Wellesbourne. Report of the National Vegetable Research Station for 1973, pp. 21–30.